Timing of angiotensin-converting enzyme inhibitor/angiotensin receptor blocker restart after cardiac surgery and postoperative outcomes: a target trial emulation

Abstract Objective Restarting angiotensin-converting enzyme inhibitors (ACEI)/angiotensin receptor blockers (ARB) after cardiac surgery is associated with better in-hospital outcomes, but evidence on when to restart after surgery is lacking. This study takes "restart timing" itself as an intervenable treatment strategy and compares the effects of four restart strategies on postoperative outcomes. Methods A target trial emulation was conducted based on the MIMIC-IV 3.1 public critical care database. Inclusion criteria: adult open cardiac surgery, preoperative use of ACEI/ARB, postoperative admission to the intensive care unit (ICU), no ACEI/ARB order in use at ICU admission (t0), and restart before discharge. Four strategies were compared: restart within 24 h, restart within 48 h, restart within 72 h, and restart from 72 h to before discharge. Twelve outcomes were analysed: all-cause death at 30, 90 and 365 days, and 9 in-hospital complication outcomes. Results Among 10 939 open cardiac surgeries, 5 496 patients used ACEI/ARB before surgery, and 1 420 patients restarted before discharge (median restart time 53.9 h). The four strategies contained 259, 664, 872 and 548 patients, respectively. There were 51 deaths within 1 year, 24 within 90 days and 12 within 30 days. Weighted mortality in the four strategies was 4.90%, 2.76%, 3.05% and 4.57% at 1 year, 1.57%, 1.32%, 1.78% and 1.80% at 90 days, and 0.40%, 0.59%, 1.13% and 0.79% at 30 days; at the two earlier time points the intervals of all 6 pairwise comparisons included 0. Compared with restart within 48 h, restart within 24 h had higher 1-year mortality (RD +2.14 percentage points, 95%CI 0.05～4.27; relative risk 1.78, 1.02～2.67; hazard ratio 1.70, 1.01～2.55; E-value 2.95), but this comparison was no longer significant after correction for 6 comparisons within the mortality family (q = 0.23). Among the 15 comparisons whose uncorrected confidence intervals did not cross 0, 8 had q values below 0.05, 3 were 0.054～0.060, and 4 were 0.09～0.23. The risk of reintubation with restart within 24 h was lower than in those restarting at other times. The risk of hyperkalaemia and liver injury was highest with restart at 72 h and later. The risk of infection, pulmonary complications and cardiovascular or cerebrovascular complications with restart within 48 h was higher than with restart at 72 h and later. Conclusions ACEI/ARB restart timing did not differ across the four strategies with respect to 30-day and 90-day mortality; the association with 1-year mortality was not robust after correction for multiple comparisons; the risk of reintubation with restart within 24 h was lower than in those restarting at other times. The risk of hyperkalaemia and liver injury was highest with restart at 72 h and later. The risk of infection, pulmonary complications and cardiovascular or cerebrovascular complications with restart within 48 h was higher than with restart at 72 h and later.